Rapid restoration of potent neutralization activity against the latest Omicron variant JN.1 via AI rational design and antibody engineering

Abstract

The rapid evolution of the viral genome has led to the continual generation of new variants of SARS-CoV-2. Developing antibody drugs with broad-spectrum and high efficiency is a long-term task. It is promising but challenging to develop therapeutic neutralizing antibodies (nAbs) through in vitro evolution based on antigen-antibody binding interactions. From an early B cell antibody repertoire, we isolated antibody 8G3 that retains its nonregressive neutralizing activity against Omicron BA.1 and various other strains in vitro. 8G3 protected ACE2 transgenic mice from BA.1 and WA1/2020 virus infection without adverse clinical manifestations and completely cleared viral load in the lungs. Similar to most IGHV3-53 antibodies, the binding sites of 8G3 and ACE2 largely overlap, enabling competition with ACE2 for binding to RBD. By comprehensively considering the binding free energy changes of the antigen-antibody complexes, the biological environment of their interactions, and the evolutionary direction of the antibodies, we were able to select 50 mutants. Among them, 11 were validated by experiments showing better neutralizing activities. Further, a combination of four mutations were identified in 8G3 that increased its neutralization potency against JN.1, the latest Omicron mutant, by approximately 1,500-fold, and one of the mutations led to an improvement in activity against multiple variants to a certain extent. Together, we established a procedure of rapid selection of neutralizing antibodies with potent SARS-CoV-2 neutralization activity. Our results provide a reference for engineering neutralizing antibodies against future SARS-CoV-2 variants and even other pandemic viruses.

Keywords: SARS-CoV-2; antibody engineering; neutralizing antibody.

Fig. 1.

In vitro neutralization activity of 8G3. (A and B) Concentration-dependent neutralization of various subtypes of pseudotyped SARS-CoV-2 by 8G3. Data with duplications are shown as means ± SD. (C and D) Concentration-dependent neutralization of authentic WA1/2020 and Omicron BA.1 by 8G3 and other EUA antibodies. Data with duplications are shown as means. (E) Concentration-dependent neutralization of several authentic SARS-CoV-2 strains by 8G3. Data with duplications are shown as means. (F) Antibody was serially diluted fivefold from 100 μg/mL and mixed with rVSV of WA1/2020, Delta, or Omicron BA.1 prior to the infection of Vero E6 cells. After 48 h, cytopathic effects in each well were examined, and medium was collected from the well with the highest concentration of virus showing positive infection for the next round of passaging. (G) Deep sequencing of rVSVs identifies escape mutation sites in the presence of antibody 2G1 or 8G3 pressurization. CPE means cytopathic effects. ND means not detected.

Fig. 2.

Therapeutic efficacy of antibody 8G3 against SARS-CoV-2 infection in transgenic mice. (A) Human ACE2 transgenic mice were challenged with 300-fold TCID50 of WA1/2020 or 2.3 × 10⁴-fold TCID50 of Omicron BA.1 viruses and then received two doses of 8G3 treatment at 4 h and 48 h postinfection. (B) Mouse survival number. (C and D) Percentage of Body weight of mice. Clinical manifestations were observed at least once daily. The data points for antibody groups are identical, resulting in complete overlap on the graph. Data with duplications are shown as means ± SD. (E and F) Clinical score of mice. The clinical well-being of mice was assessed based on a 1 to 4 grading system. Data with duplications are shown as means ± SD. (G and H) Viral titer in lung and brain tissues on day 4 and day 7; W represents WA1/2020, and O represents BA.1. Data are shown as mean ± SD. Statistical differences were determined by one-way ANOVA, in which we set the data of body weight of dead mice in (C) and (D) to 0, and the clinical scores of dead mice in (E) and (F) to 4 (the maximum value). *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Fig. 3.

Cryo-EM structural analysis of the 8G3 antibody and Omicron BA.1 RBD complex. (A) Domain-colored cryoelectron microscopy structures of the SARS-CoV-2 S ectodomain trimer and 8G3 fragment complexes. (B) Cryo-EM structure of 8G3 and the complex with BA.1 RBD. (C) Amino acid residues of the RBD that interact with ACE2, the heavy chain of 8G3, and the light chain of 8G3. (D–I) Details of the binding of BA.1 RBD (cyan) to the heavy (D–H) and light (I) chains of 8G3. (F) Density map of residues 455 to 456 of RBD and the amino acids on the antibody heavy chain within 5 Å of these residues. The density map is displayed at a contour level of 0.6. In this study, the amino acid residues of the antibodies were numbered in sequence order starting from the variable region without considering the regional and structural characteristics of the antibodies.

Fig. 4.

Computational recovery of 8G3’s neutralization activity against JN.1. (A) Schematic of the optimization steps using a rational combination of computational methods. (B) Single-point mutations that improved 8G3’s neutralization against JN.1. (C) Mutation combinations that enhanced 8G3’s neutralization against JN.1. (D and E) Comparison of the concentration-dependent neutralization effects of the mutants that improved 8G3’s neutralization activity against JN.1. Data are presented as mean ± SD from repeated experiments.

Fig. 5.

Structural analysis of the mutated 8G3 antibody with JN.1 RBD. (A) The positions of S31M, G101F, Y91M, and N96V mutations in the JN.1 RBD–8G3 complex. (B) G101F effectively counteracts L455S-induced antibody escape from 8G3. (C) Mutation at position 96 in the light chain to a hydrophobic amino acid stabilizes the antigen–antibody complex. The residue highlighted in forest green is ASN405. Using this as a reference, it is evident that the volume of TYR is larger than that of VAL. (D) S31M mutation enhances hydrophobic interactions between the heavy chain of the antibody and JN.1 RBD. (E) Y91M mutation promotes tighter binding between the heavy and light chains of the antibody. Binding models of JN.1 RBD–8G3 were generated by Rosetta based on cryoelectron microscopy results.